The long-term goal of this study is the development of preventive measures for cancer of breast. The principal specific aims are to relate mammographic features of breast tissue to the dietary intakes of total fat, saturated fat, carotenoids and fiber, and to serum levels of beta-carotene, other carotenoids, retinol and vitamin E. The study will be based on women attending a state-wide screening program for breast cancer in Rhode Island. A total of 2000 women will be identified who are 40-59 years old at screening, who do not report symptoms of breast disease, and who do not have mammographic findings that are suspicious of cancer. Subjects will be interviewed at home with respect to dietary history and other known or suspected risk factors for breast cancer. The dietary history will use the food frequency method supplemented with food models to help estimate portion size. The data will be used to estimate intake of fats, carotenoids, fiber and other nutrients in the year preceding the interview. Subjects will also be asked for a blood specimen. These will be analyzed by high-performance liquid chromatography for retinol, fractionated carotenoids, and alpha- and gamma-tocopherol. Mammograms will be evaluated by a single reviewer without reference to other data on subjects. Features to be evaluated include Wolfe's parenchymal pattern, the radiologic components of the parenchymal pattern, and the breast size. Relations between dietary intakes, serum measurements and mammographic characteristics will be analyzed by use of tabular and regression methods.